extremevariantfilter 0.0a3

A set of tools to aid in the identification of false positive variants in Variant Call Files.

# ExtremeVariantFilter

Extreme Variant Filter is a set of tools developed to aid in the identification of false positive variants in Genomic Variant Call Files based on XGBoost.

### Functions

__apply_filter__

Usage:

apply_filter (–vcf STR) (–snp-model STR) (–indel-model STR) [–verbose]

Description:

Apply models generated by train_model to a VCF.

Arguments:

--vcf STR

VCF to be filtered

--snp-model STR

Model for applying to SNPs

--indel-model INT

Model for applying to InDels

Options:

-h, --help

Show this help message and exit.

-v, --version

Show version and exit.

--verbose

Log output

Examples:

apply_filter –vcf <table> –snp-model <snp.pickle.dat> –indel-model <indel.pickle.dat>

__train_model__

Usage:

train_model (–true-pos STR) (–false-pos STR) (–type STR) [–out STR] [–njobs INT] [–verbose]

Description:

Train a model to be saved and used to filter VCFs.

Arguments:

--true-pos STR

Path to true-positive VCF from VCFeval or comma-seperated list of paths

--false-pos STR

Path to false-positive VCF from VCFeval or comma-seperated list of paths

--type STR

SNP or INDEL

Options:

-o, --out <STR>

Outfile name for writing model [default: (type).filter.pickle.dat]

-n, --njobs <INT>

Number of threads to run in parallel [default: 2]

-h, --help

Show this help message and exit.

-v, --version

Show version and exit.

--verbose

Log output

Examples:

train_table –true-pos <path/to/tp/vcf(s)> –false-pos <path/to/fp/vcf(s)> –type [SNP, INDEL] –njobs 20

### Install

To install and run EVF simply type:

pip install extremevariantfilter

Alternatively, clone this repo and build using the following commands:

git clone https://github.com/stLFR/extremevariantfilter.git cd extremevariantfilter pip install .

### stLFR Paper Results

If you’d like to use this tool to corroborate the results from the [stLFR Paper on Bioarxiv](https://www.biorxiv.org/content/early/2018/05/17/324392.1) paper, the models used for variant filtering are available within the models/ directory. In order to get identical results, after installation, use the command pip install -r requirements.txt from within this directory to ensure your environment matches the one we used for our results. Different versions of certain packages will result in variable results.